Skip to content
English

Switch theme

Discover▾

Not sure where to start? This week’s hottest software, models and stacks in three minutes.

AI Apps▾

AI software people use: from assistants like ChatGPT to tools for coding, video and workflows, picked by purpose.

For Agents▾

Software for AI: tools that let your agent (Claude Code, Codex, Cursor…) browse, read docs, run code and remember.

Models▾

Choose models, estimate cost, self-host: prices, rankings, benchmarks, free tiers and compute.

Solutions▾

A goal or an error message? Ready-made stacks, steps and fixes by scenario.

Learn▾

A systematic course from zero to the frontier, with an algorithm library, must-read papers and tutorials.

Frontier▾

Daily AI news, weekly highlights, the latest papers and how AI is changing every discipline.

// Open-model VRAM table · family

DeepSeek: local deployment & VRAM

34 open-weight models from DeepSeek (42 official repos incl. quantised builds). At 4-bit they range from 1.9 GB (DeepSeek-R1-Distill-Qwen-1.5B) to 1.10 TB (DeepSeek-V4-Pro-0813).

Filter by my hardware → VRAM calculator

official orgs: deepseek-ailatest 2026-10-17memory = estimate

01Text LLM (27)

ModelParamsContextBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
DeepSeek-V4-Pro-08131.65T 1M 3.63 TB1.10 TB 8 × B200 (1,440GB)— Commercial OK2026-08-13
DeepSeek-V4-Flash-0731304.2B 1M 670 GB203 GB 4 × RTX PRO 6000 (384GB)Mac 512GB Commercial OK2026-07-31
DeepSeek-V4-Flash284B / 13B 1M 625 GB190 GB 4 × RTX PRO 6000 (384GB)Mac 512GB Commercial OK2026-04-22
DeepSeek-V4-Pro1.6T / 49B 1M 3.52 TB1.07 TB 8 × B200 (1,440GB)— Commercial OK2026-04-22
DeepSeek-V3.2685.4B 160K 1.51 TB458 GB 8 × H100 (640GB)— Commercial OK2025-12-01
DeepSeek-V3.2-Speciale685.4B 160K 1.51 TB458 GB 8 × H100 (640GB)— Commercial OK2025-11-28
DeepSeek-V3.2-Exp-Base685.4B 160K 1.51 TB458 GB 8 × H100 (640GB)— Commercial OK2025-09-29
DeepSeek-V3.1-Base671B / 37B 160K 1.48 TB449 GB 8 × H100 (640GB)— Commercial OK2025-08-19
DeepSeek-R1-0528-Qwen3-8B8.19B 128K 19.7 GB7.2 GB 8GB GPUMac 16GB Commercial OK2025-05-29
DeepSeek-R1-0528684.5B 160K 1.51 TB458 GB 8 × H100 (640GB)— Commercial OK2025-05-28
DeepSeek-V3-0324684.5B 160K 1.51 TB458 GB 8 × H100 (640GB)— Commercial OK2025-03-24
DeepSeek-R1671B / 37B 160K 1.48 TB449 GB 8 × H100 (640GB)— Commercial OK2025-01-20
DeepSeek-R1-Distill-Llama-8B8.03B 128K 19.2 GB6.9 GB 8GB GPUMac 16GB Commercial OK2025-01-20
DeepSeek-R1-Distill-Llama-70B70.6B 128K 158 GB50.2 GB 2 × RTX 5090 (64GB)Mac 128GB Commercial OK2025-01-20
DeepSeek-R1-Distill-Qwen-1.5B1.78B 128K 4.6 GB1.9 GB 8GB GPUMac 16GB Commercial OK2025-01-20
DeepSeek-R1-Distill-Qwen-7B7.62B 128K 17.7 GB6.0 GB 8GB GPUMac 16GB Commercial OK2025-01-20
DeepSeek-R1-Distill-Qwen-14B14.8B 128K 34.6 GB12.0 GB 16GB GPUMac 24GB Commercial OK2025-01-20
DeepSeek-R1-Distill-Qwen-32B32.8B 128K 74.7 GB24.5 GB RTX 5090 32GBMac 64GB Commercial OK2025-01-20
DeepSeek-R1-Zero671B / 37B 160K 1.48 TB449 GB 8 × H100 (640GB)— Commercial OK2025-01-20
DeepSeek-V3671B / 37B 160K 1.48 TB449 GB 8 × H100 (640GB)— Licence not stated2024-12-25
DeepSeek-V2.5-1210235.7B 160K 520 GB158 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial with conditions2024-12-10
DeepSeek-V2.5235.7B 160K 520 GB158 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial with conditions2024-09-05
DeepSeek-Coder-V2-Instruct-0724236B / 21B 160K 520 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial with conditions2024-09-05
DeepSeek-V2-Chat-0628235.7B 160K 520 GB158 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial with conditions2024-07-18
DeepSeek-Coder-V2-Instruct236B / 21B 160K 520 GB159 GB 4 × RTX PRO 6000 (384GB)Mac 256GB Commercial with conditions2024-06-14
DeepSeek-Coder-V2-Lite-Instruct16B / 2.4B 160K 36.0 GB11.4 GB 16GB GPUMac 24GB Commercial with conditions2024-06-14
DeepSeek-V2-Lite16B / 2.4B 160K 36.0 GB11.4 GB 16GB GPUMac 24GB Commercial with conditions2024-05-15

02Vision-language (7)

ModelParamsContextBF164-bitSmallest GPU (4-bit)Mac (4-bit)LicenceReleased
DeepSeek-V4.1-Flash-1017748B / 16B 1M 1.65 TB499 GB 8 × H100 (640GB)— Commercial OK2026-10-17
DeepSeek-V4.1-Flash748B / 16B 1M 1.65 TB499 GB 8 × H100 (640GB)— Commercial OK2026-09-10
DeepSeek-OCR-23.39B 8K 8.5 GB3.3 GB 8GB GPUMac 16GB Commercial OK2026-01-27
DeepSeek-OCR3.34B 8K 8.3 GB3.2 GB 8GB GPUMac 16GB Commercial OK2025-10-17
deepseek-vl227.5B — 61.0 GB18.8 GB RTX 4090 24GBMac 36GB Commercial with conditions2024-12-13
deepseek-vl2-small16.1B / 2.8B 4K 37.8 GB13.1 GB 16GB GPUMac 24GB Commercial with conditions2024-12-13
deepseek-vl2-tiny3.37B / 1B 4K 8.4 GB3.2 GB 8GB GPUMac 16GB Commercial with conditions2024-12-13

03Other families

Qwen · IBM Granite · NVIDIA Nemotron · Gemma · Mistral / Magistral / Devstral · GLM / CogVideoX (Z.ai) · Ai2 OLMo / Molmo · Ant Ling / Ring / Ming · Liquid LFM · MiniCPM / VoxCPM · Tencent Hunyuan / Hy · Phi · Llama · Meituan LongCat · Jina embeddings & rerankers · Wan · Kimi · Cohere Command / Aya · Xiaomi MiMo · FLUX · Microsoft VibeVoice / Fara / Harrier · InternLM / Intern-S · StepFun Step · MiniMax · Baidu ERNIE · NVIDIA Parakeet / Canary speech · Stable Diffusion / Stable Audio · Yi (01.AI) · ByteDance Seed · SmolLM / SmolVLM · BGE embeddings & rerankers · LTX Video · FunAudioLLM speech · gpt-oss · poolside Laguna · Whisper · Baichuan · Thinking Machines Inkling · xAI Grok · rednote dots

Totals include 10% + 0.5GB overhead and one 8K-token KV cache for text models. Estimates only. Method →

Data updated: Self-host open models(UTC+8)